Horizontal and vertical crossover of sine cosine algorithm with quick moves for optimization and feature selection

Abstract The sine cosine algorithm (SCA) is a metaheuristic proposed in recent years that does not resort to nature-related metaphors but explores and exploits the search space with help of two simple mathematical functions cosine. SCA has fewer parameters structure widely used various fields. However, it tends fall into local optimality because have well-balanced exploitation exploration phase. Therefore, this paper, new, improved (QCSCA) improve performance by introducing quick move mechanism crisscross adaptively improving one parameters. To verify effectiveness QCSCA, comparison experiments some conventional algorithms, advanced variants are conducted on IEEE CEC2017 CEC2013. experimental results show significant improvement convergence speed ability jump out optimum QCSCA. scalability verified benchmark function. In addition, QCSCA applied 14 real-world datasets from UCI machine learning database for selecting subset near-optimal features, still very competitive feature selection (FS) compared similar algorithms. Our analysis an effective method solving global optimization problems FS problems.